Dear Tirumala Marri,
In message f5b257a2367890d26637726f5c8e1dea@mail.gmail.com you wrote:
On the other hand, this seems to be a 464 core, so should we not print 464 here?
[Marri] Interesting point, For all other processors we were printing SoC name not the CPU core name. APM82xxx is SoC name.
OK.
Hm.., the whole CPU ID code in "arch/powerpc/cpu/ppc4xx/cpu.c" has become a mess.
Stefan:
is there any deeper logic for the repeated #ifdef's there, when we finally do a switch on the PVR anyway?
Instead of painstakingly concatenating "AMCC PowerPC 4" with "05" and later "GP Rev. B" we could shrink this into a plain
case PVR_405GP_RB: puts("AMCC PowerPC 405GP Rev. B"); break; Or am I missing something?
Let's get rid of all these nested #ifdef's...
diff --git a/include/ppc440.h b/include/ppc440.h index c807dda..b5c1832 100644 --- a/include/ppc440.h +++ b/include/ppc440.h
Hm... you are adding a number of large blocks of code, all #ifdef'ed.
This is ugly, difficult to read and difficult to maintain.
Why not adding a new file include/apm82xxx.h instead?
[Marri] If we create new file it would cause too many changes in the common files. When more and more APM8xxxx processors with new registers definitions being released At that point we can create separate file. Please suggest ?
If there are plans for more and more APM8xxxx processors with new registers definitions to be released we should not create a mess now and later go into efforts to clean it up again, but start with a separate register definitions file apm82xxx.h now.
Best regards,
Wolfgang Denk